Abstracts

Rationale: Previous research demonstrated that memory associated fMRI activation induced by a simple visuo-spatial recall fMRI task (Hometown Walk) correlated with postsurgical non-verbal memory outcome in patients with right-sided temporal lobe epilepsy (TLE, see Janszky et al., 2007, Epilepsia, 46, 244-250). Patients with left-sided TLE are specifically at risk of verbal memory decline after epilepsy surgery. In the current study, we investigated whether left-sided TLE patients with presurgically intact but postsurgically reduced verbal memory showed differing activation patterns on presurgical fMRI using the Howmtown Walk compared to left-sided TLE patients without postsurgical memory loss. We also analyzed whether visuo-spatial memory activations were associated with postsurgical verbal memory changes. Methods: We studied 16 left-sided TLE patients with hippocampus sclerosis with presurgically unimpaired verbal memory (percentiles >16 on the direct and delayed recall trials of the Verbal Learning and Memory Test, VLMT). Seven patients had postsurgical verbal memory decline (>1 SD reduction on the delayed recall trial of the VLMT), nine did neither deteriorate nor improve. Prior to mesiotemporal resection all patients were investigated with the fMRI version of the Hometown Walk. In this blocked fMRI task patients had to mentally navigate through their everyday environment and recall as many details as possible. Results: The whole patient group exhibited left- and right-sided patterns of hippocampal-parahippocampal and prefrontal memory activations on the Hometown Walk (p<.001 corrected). Patients with memory loss did not show stronger activations compared to the postsurgically unimpaired patients in any brain region. In contrast, those patients with preserved postsurgical verbal memory exhibited stronger activations within the left fusiform gyrus (x=-48, y=-51, z=-18, k=28, t=5.3, p=<.001) compared to the patients with memory loss. However, we did not find correlations between postsurgical changes in verbal memory and presurgical activations associated with the Hometown Walk, neither for the activation strength in the cluster that differentiated between the two groups nor for hippocampal or prefrontal activations. Conclusions: Stronger activations within the left fusiform gyrus in those patients without postsurgical memory reductions might reflect successful intrahemispheric reorganization. However, we did not observe direct associations between presurgical visuo-spatial memory activations and postsurgical verbal memory alterations. Thus, the use of a visuo-spatial fMRI memory task seems not to be helpful to directly predict postsurgical verbal memory outcome in patients with left-sided TLE.